Proliferative diabetic retinopathy with macular oedema in a woman who is 28 weeks pregnant

This is treatment-requiring proliferative diabetic retinopathy with vitreous haemorrhage and centre-involving oedema in a pregnant patient, which is one of the few situations where the usual 'anti-VEGF first' reflex must be overridden.

Anti-VEGF agents are systemically absorbed after intravitreal injection, and VEGF is central to placental and fetal vascular development. Human data are limited to case reports, and the pregnancy safety profile is unknown for all three agents; aflibercept has the longest systemic half-life because of its Fc portion and would be the least attractive choice. There is also a theoretical concern that rapid VEGF suppression could affect placental function, particularly in a woman with hypertension.

My default would be panretinal photocoagulation now, delivered in two or three sessions with adequate analgesia, treating the areas not obscured by haemorrhage. PRP is the established, systemically inert treatment and is not contraindicated in pregnancy. Where haemorrhage obscures the view, treat the visible retina first and complete the treatment as the vitreous clears; if a dense non-clearing haemorrhage prevents adequate PRP and the fellow eye or the perfusion status demands it, vitrectomy with endolaser is feasible in pregnancy with left lateral tilt after 20 weeks and avoidance of hypotensive anaesthesia.

Macular oedema in pregnancy is often self-limiting after delivery. Given the competing risk of progression, I would use PRP as the primary tool, monitor the macula closely, and reserve anti-VEGF for sight-threatening progression after a documented multidisciplinary discussion with obstetrics and maternal medicine, and ideally after delivery.

Several systemic threads need to be pulled together, because diabetic retinopathy in pregnancy is a systemic disease with an ocular endpoint.

1. Blood pressure of 148/92 with a rising trend requires obstetric input; pre-eclampsia can present with visual symptoms and must stay on the differential even though current testing is negative. Good blood pressure control reduces progression, but aggressive lowering can also reduce ocular perfusion, so it should be coordinated rather than driven by the eye clinic.
2. Rapid glycaemic improvement from 10.4% to 8.9% is a recognised cause of early worsening of retinopathy; the current florid proliferation is consistent with it. This is an argument for tighter surveillance, not for loosening control - the long-term benefit of control is clear, but she needs review every 4 to 8 weeks until delivery.
3. Delivery planning: mode and timing are obstetric decisions, but the team should know she has active proliferation. Prolonged Valsalva in the second stage, and any hypertensive emergency, raise the risk of vitreous haemorrhage. This is information to share, not a reason to mandate caesarean section.
4. The left eye: severe non-proliferative disease at 28 weeks carries a high risk of progressing to proliferation before term. Review monthly and treat promptly if neovascularisation appears.

Postpartum, she should be reviewed within 4–6 weeks, when the oedema can be reassessed and anti-VEGF therapy considered; discuss lactation with her explicitly, since data are limited and this influences timing. Also arrange her routine diabetic retinal screening pathway after the pregnancy so she does not fall out of surveillance.

This is a non-clearing dense vitreous haemorrhage in a pregnant woman with active proliferation - the decision is no longer about lasers and injections but about whether to operate before delivery.

Options:

• Observe to delivery and reassess postpartum. Reasonable if there is no evidence of tractional retinal detachment and the fellow eye is stable; many diabetic vitreous haemorrhages clear over weeks to months, and surgery in pregnancy carries anaesthetic and positioning complexities. The disadvantage is that a tractional detachment may develop unseen behind the haemorrhage.
• Ultrasound is the key next test: it will show whether the retina is attached, whether there is a tractional detachment involving the macula, and whether there is a posterior vitreous detachment. It is safe in pregnancy and should be done at this visit, not deferred.
• Pars plana vitrectomy before delivery is indicated if ultrasound shows a macula-threatening tractional detachment, or if both eyes are severely affected. It can be performed in the third trimester with left lateral tilt after 20 weeks, thromboprophylaxis, and avoidance of hypotensive anaesthesia; the obstetric team should be involved in planning and in fetal monitoring.

My advice to her: ultrasound today; if the macula is attached and there is no tractional detachment, plan review every two weeks, deliver with obstetric guidance, and reassess at 4–6 weeks postpartum, with vitrectomy if the haemorrhage has not cleared or if traction develops. Counsel her that vision may be poor for some weeks but that the anatomical outcome of vitrectomy for non-clearing diabetic haemorrhage is generally good, and that the priority now is protecting the macula and completing the pregnancy safely.